Multivariate data reduction techniques have recently become available through statistical computer packages. Through both factor analysis and cluster analysis, variables and cases respectively may be aggregated. These techniques were applied to a sample of schizophrenic patients who had received the Wisconsin Card Sort. In the latter study, data that were based on means were employed. However, the approach might have obscured distinct subtypes. Therefore, we attempted to cluster patients based on performance. However, we found that even small changes in the statistical programs led to major differences in the solutions we derived.